Unify execution layer endpoints (#3214)

## Issue Addressed

Resolves #3069 

## Proposed Changes

Unify the `eth1-endpoints` and `execution-endpoints` flags in a backwards compatible way as described in https://github.com/sigp/lighthouse/issues/3069#issuecomment-1134219221

Users have 2 options:
1. Use multiple non auth execution endpoints for deposit processing pre-merge
2. Use a single jwt authenticated execution endpoint for both execution layer and deposit processing post merge

Related https://github.com/sigp/lighthouse/issues/3118

To enable jwt authenticated deposit processing, this PR removes the calls to `net_version` as the `net` namespace is not exposed in the auth server in execution clients. 
Moving away from using `networkId` is a good step in my opinion as it doesn't provide us with any added guarantees over `chainId`. See https://github.com/ethereum/consensus-specs/issues/2163 and https://github.com/sigp/lighthouse/issues/2115


Co-authored-by: Paul Hauner <paul@paulhauner.com>

lcli/src/eth1_genesis.rs

@@ -1,7 +1,7 @@
 use clap::ArgMatches;
 use environment::Environment;
 use eth2_network_config::Eth2NetworkConfig;
-use genesis::{Eth1Config, Eth1GenesisService};
+use genesis::{Eth1Config, Eth1Endpoint, Eth1GenesisService};
 use sensitive_url::SensitiveUrl;
 use ssz::Encode;
 use std::cmp::max;
@@ -35,11 +35,12 @@ pub fn run<T: EthSpec>(
 
     let mut config = Eth1Config::default();
     if let Some(v) = endpoints.clone() {
-        config.endpoints = v
+        let endpoints = v
             .iter()
             .map(|s| SensitiveUrl::parse(s))
             .collect::<Result<_, _>>()
             .map_err(|e| format!("Unable to parse eth1 endpoint URL: {:?}", e))?;
+        config.endpoints = Eth1Endpoint::NoAuth(endpoints);
     }
     config.deposit_contract_address = format!("{:?}", spec.deposit_contract_address);
     config.deposit_contract_deploy_block = eth2_network_config.deposit_contract_deploy_block;